Combined Stabilizing of the Solid–Electrolyte Interphase with Suppression of Graphite Exfoliation via Additive-Solvent Optimization in Li-Ion Batteries

Propylene carbonate (PC) is a promising solvent for extending the operating temperature range for lithium-ion batteries (LIBs) because of its high dielectric constant and wide temperature range stability. However, PC can cause graphite exfoliation through cointercalation, leading to electrolyte decomposition and subsequent irreversible capacity loss. This work reports the formulation of a ternary electrolyte with the introduction of an inorganic salt additive, potassium hexafluorophosphate (KPF6), to address the aforementioned concerns. We demonstrate the cumulative effect of solvent and additive on delivering multiple performance benefits and safety of the battery. The faster diffusion rate of K + solvation shell decreases the rate of PC decomposition, thereby reducing its cointercalation. Additionally, the optimum concentration of KPF6, i.e., 0.1 M constructs a robust and insoluble LiF-rich electrode/electrolyte interphase, further suppressing graphite exfoliation and Li dendrite formation. The stable cyclability is achieved by enhanced Li + transportation through the LiF-rich interphase, enabling an exfoliation-free and dendrite-free graphite anode in the ternary electrolyte.

Incremental capacity plot for E-20PC (without KPF 6 additive) and E-20PC-0.001Mshow crowded peaks during charging with no reduction peak during discharging in Figure S1a.This signifies that Li + ions are consumed through electrolyte decomposition, delivering poor specific capacities in Figure 1a.Moreover, well defined two oxidation and two corresponding reduction peaks are monitored with the increase in additive content to 0.1M, which perfectly aligns with the dQ/dV vs V peak of commercial RD281 electrolyte 1 , confirming Li + lithiation and delithiation through graphite layers has occurred in E-20PC-0.1Melectrolyte.
In Figure S1b, E-10PC shows two sharp oxidation and reduction dQ/dV vs V peaks without the additive, demonstrating Li + intercalation and de-intercalation through the graphite anode, unlike the E-20PC electrolyte.It is observed that the oxidation peak attributed to Li + intercalation becomes sharper and eventually shifts towards the left with increasing in additive content to 0.1M in 10PC electrolytes, ensuring reduced polarisation compared to other electrolytes.

S3
Moreover, Figure S1 denotes that the addition of 0.1M KPF 6 reduces electrolyte decomposition and polarisation in 20PC and 10PC respectively, improving cells performance upon cycling.For a clearer understanding, a comparison of specific discharge capacities with respect to the concentration of KPF 6 additive for 20PC and 10PC are presented in Figure S4a and b.It is observed that the specific discharge capacity is doubled when the PC content is decreased from 20 vol% to 10 vol% in the very first cycle (Figure S4a) and maintained the similar capacity difference even on 100 th cycle (Figure S4b).This demonstrates that the specific discharge capacities are majorly affected by the PC content in the electrolyte.High PC content i.e., 20 vol% produces more soluble decomposition products (lithium propylene dicarbonate, LPDC) 2 , hence, builds a loose SEI film on graphite anode.Therefore, the probability of graphite exfoliation via PC co-intercalation is high in 20PC electrolyte, leading to the poor electrochemical performance [3][4][5] compared to 10PC based electrolytes.

S4
With introduction of KPF 6 additive into the electrolyte, an increase in specific discharge capacities is obtained in both 20PC and 10PC electrolytes.In addition to this, the difference in specific discharge capacities (between cells cycled with 20PC and 10PC electrolytes) is decreased, when KPF 6 concentration is increased to 0.01M.Lastly, the difference in capacities is reduced to ~ 1 mAh/g NMC when 0.1M KPF 6 is added to the 20PC and 10PC electrolytes.The S7 similar trend is observed on the 100 th cycle, presented in Figure S4b, indicating the stable cyclability of the cell using modified electrolytes.XRD diffraction refinement plots of graphite anode cycled with RD281, E-10PC, E-10PC-0.1Melectrolytes are presented Figure S5.For a comparison, the refinement of pristine graphite anode is carried out in Figure S5a.The refinement provides the c-lattice spacing of graphite, from which interlayer d-spacing is obatinaed.The refinement results in c-lattice parameter of 6.71798(280) Å for E-20PC in electrolyte Figure S5e, suggesting the expansion of graphite layers 6 .However, the c-lattice parameter for E-10PC, E-20PC-0.1Mand E-10PC-0.1Mare 6.70803(8) Å, 6.70836(120) Å and 6.707947(85) Å respectively, which is close to the value obtained from the pristine graphite i.e., 6.70917(358) Å.This confirms the presence of relatively stable SEI film in E-20PC-0.1Mand E-10PC-0.1Melectrolytes, which suppresses in PC co-intercalation via subsequent graphite exfoliation (due to 0.1M KPF 6 electrolyte additive).The impact of KPF 6 additive concentrations on graphite anode in 20PC electrolytes is presented in Figure S6.The exfoliated graphite layers are clearly evident in Figure S6a and e. PC solvated Li + ions co-intercalates into the graphite layers and further decomposes to gaseoue products, which acts as a source of graphite exfoliation 7 .The exfoliation appears to be reduced with the incorporation of KPF 6 additive in Figure S6b and c.Ultimately, with 0.1M concentration, graphite exfoliation is fully suppressed, which aligns with the XRD results in Figure 3c and d The impact of KPF 6 additive concentrations on graphite anode in 10PC electrolyte is presented in Figure S7.A large cluster of Li dendrites is observed (presented as red arrow) on graphite surface in E-10PC (Figure S7a).This deposition of Li dendrite is also detected electrochemically in Figure 2e, when graphite potential falls below 0 V vs Li/Li + .Li deposition is seemingly reduced, when the additive concentration increases from 0.001M to 0.01M (Figure S7b and c).Lastly, 0.1M KPF 6 in E-10PC-0.1Melectrolyte, Li dendrites appears to be prevented in Figure S7d, which is consistent with the electrochemical cycling result presented in Figure 2e.Moving on to 0.15M KPF 6 , four dQ/dV vs V peaks are observed identical to E-20PC-0.1Mand E-10PC-0.1Mduring 1 st cycle in Figure S8a.However, dQ/dV vs V plot for 100 th cycle in Figure S8b is found to have single peak both in charging and discharging processes especially in E-20PC-0.15M.Similar behaviour have been reported previously 1 with KPF 6 modified electrolytes.Metallic potassium deposition is believed to be the cause of this behaviour.

S12
Table S2.Potassium deposition potential with respect to the concentration of KPF 6 additive in the electrolytes.Table S2 shows the equillibrium potassium deposition potential with respect to the concentrations of KPF 6 additive in the electrolytes.This denotes that potassium metal deposition starts when graphite anode achieves these potentials.

Concentration of KPF
The potassium deposition potential for optimum 0.1M KPF 6 concentration is 0.062 V.
However, graphite potentials achieved by 0.1M KPF 6 additive (in E-20PC-0.1Mand E-10PC-0.1M)are 0.07 V and 0.08 V (See, Figure S9b), ensuring no metallic potassium deposition on graphite anode.This confirms that potassium in ionic form K + assists in building stable SEI in 20PC and 10PC based electrolytes.Three electrode El-cells were assembled for monitoring graphite anode potnetial upon cycling.
A zoomed-in plot is presented in Figure S9a.The specific discharge capacity of E-20PC is significantly lower compared to other electrolytes cause by excessive electrolyte decomposition via graphite exfoliation, as shown in Figure 1a, 2a and, 3c.In E-10PC, a negative graphite potential (-0.02V) is measured, indicating the deposition of Li metal due to the polarisation developed in the cell.However, a significantly high specific discharge capacity of ~ 150 mAh/g NMC is observed when 0.1M KPF 6 is added into 20PC and also 10PC electrolytes.E-20PC-0.1Machieves a positive graphite potential of ~ 0.07 V by suppressing graphite exfoliation (see, Figure 3c and d).Similarly, E-10PC-0.1Machieves a positive potential of ~ 0.08 V by inhibiting Li metal deposition (also shown in Figure 3h).In addition, graphite potential for E-20PC-0.1Mand E-10PC-0.1Melectrolytes is higher than the potassium deposition potential (0.062 V for 0.1M KPF 6 ), indicating no sign of potassium deposition on graphite anode.This reveals the enhanced diffusion of the K + solvation shell (potassium in ion form), due to the smaller stokes radius (compared to Li + ) 8,9 assists in constructing stable SEI S14 to inhibit graphite exfoliation and polarisation.Therefore, 0.1M KPF 6 is proved as an optimum additive concentration for PC based electrolytes.The SIMS maps of 39 K + and 41 K + ion species in positive ion modes are presented in Figure S11a and b.It is observed that 7 Li + and 19 F -secondary ion fragments (shown in Figure 5g) dominate the graphite surface in positive and negative ion modes through electrolyte decomposition, compared to 39 K + and 41 K + ion species.The mass spectra were repeated for several cycles at the same location on the graphite surface to collect ion species associated with potassium fluoride (KF).Figure S12a shows the positive ion mode mass spectra, where two peaks are detected at 97 and 99 amu.This could be attributed to 39 K 2 F + (97 amu) and 39 K 41 KF + (99 amu).Similarly, tiny peaks observed at 155 amu and 157 amu could be associated with 39 K 3 F 2 + (155 amu) and 39 K 2 41 KF 2 + (157 amu).However, the intensity ratios ( 97 I/ 99 I and 155 I/ 157 I) of 39 K 2 F + (97 amu)/ 39 K 41 KF + (99 amu) and 39 K 3 F 2 + (155 amu)/ 39 K 2 41 KF 2 + (157 amu) does not match with the relative abundance ratio of potassium isotopes ( 39 K/ 41 K = 13.8) 11, indicating the absence of KF in the SEI layer.
Similarly, on negative ion mode in Figure S12b, no peak is detected at 77 amu corresponding to 39 KF 2 -ion fragment, confirming the absence of KF in the SEI layer.

Figure S1 .
Figure S1.dQ/dV vs V plot of graphite | NMC 622 full cell with various concentrations of KPF 6 additive in (a) 20PC and (b) 10PC based electrolytes.

Figure S4 .
Figure S4.Comparison of specific discharge capacities of NMC 622 | graphite full cell with respect to

Figure S5 .
Figure S5.XRD diffraction refinement plots of (a) Pristine graphite, and graphite anodes cycled with

Figure S9 .
Figure S9.(a) Three electrode EL-cell graphite potential (graphite | Li) comparison of without additive

Figure S10 .
Figure S10.XPS K 2s spectra of graphite anode cycled with E-10PC-0.1M-confirming the presence of potassium element on the anode surface.

Figure S11 .
Figure S11.SIMS maps of secondary ion fragments of (a) 39 K + and (b) 41 K + on graphite surface cycled

Figure S12 .
Figure S12.SIMS (a) positive ion and (b) negative ion mode repeated mass spectra for a number of